prisons, a group identiÞed in this study as having a
higher likelihood of TB. At entry, whether during the
day or at night, prison staff can ask ÔDo you have
cough?Õ and if they respond afÞrmatively, can then
ask ÔHow many days have you had a cough?Õ Prisoner-
staff relationships, trust and conÞdentiality may be
important issues, but may be balanced by staff ability
to see everyone at entry.
The overall cost using the WHO scoring system
may be higher, given the large number of smears need-
ing evaluation; in this study, the number referred for
microscopy would have been reduced by 76% using
the ISTC standard. A formal cost-effectiveness analy-
sis, accounting for cost of laboratory technician time,
is needed. If possible, such an analysis would also ac-
count for the possible reduction in quality associated
with one technician performing a large number of
smears. The concern of reduced detection of cases by
22% using the ISTC standard rather than the WHO
tool may be diminished if missing cases are identiÞed
later by passive case detection, as prisoners with symp-
toms present at the ambulatory unit to be seen by
health care workers. In resource-limited settings, the
ISTC criterion for screening patients may be a better
strategy than the WHO tool for initially Þnding TB
suspects in prisons. Kimerling and Kluge urge the iden-
tiÞcation of acceptable entry screening as part of a com-
prehensive control strategy,3 and we suggest that the
ISTC standard administered by prison staff is a good
candidate for further study.
Our Þndings are limited because they were mea-
sured under routine programme conditions. We were
unable to perform sputum smears or chest X-ray
among prisoners who were not TB suspects, and TB
prevalence may therefore have been underestimated.
The WHO recommends that persons in the general
population should have at least two positive sputum
smears to be diagnosed with smear-positive TB.7,8
H owever, in this study, one positive smear was con-
sidered diagnostic, leading to possible errors. In most
countries, the prevalence of TB in prisoners is sub-
stantially greater than in the general population. The
positive predictive value of one positive smear is prob-
ably sufÞciently high to justify a diagnosis of TB, but
comparisons with other screening programmes using
two positive smears are not possible. Furthermore,
not knowing the HIV status may have resulted in an
underestimation among prisoners with HIV and TB
disease whose smears were negative. Furthermore,
1735 TB suspects were not completely evaluated, be-
cause most were transferred to other prisons. More
effort should be devoted to the issue of mobility among
the prison population. These prisoners should be a pri-
ority because they are at high risk for the disease and
they are frequently moved between prisons. Lastly, this